Open anelles opened 3 years ago
This issue is posted here to document the discussion, so please all contribute, especially @cg-laser who suggested this merger.
the plan would be to keep NuRadioReco independent in the sense that it would still be a NuRadioReco
folder in the NuRadioMC repository and you would still do the same imports via from NuRadioReco.utilities import units
for example. But all code development would be in the NuRadioMC repo.
The main reason is that it is super cumbersome to keep new development (and the tests) in sync between MC and Reco.
At the moment, NuRadioMC does not have more (hard) dependencies than NuRadioReco, so installation should be as simple as before, and probably a independent checkout of only the NuRadioReco part would be possible.
Also in the future, we should keep the base version lightweight and require more dependencies only on demand (e.g. proposal for the muon/tau propagation, or RadioPropa for a specific propagation module).
Hi, when described as above by @cg-laser I dont see any problem. It would be important for us to keep NuRadioReco independent itself (with or without an independent checkout) in the future.
@fschlueter do you have any unmerged branches that you'd want to merge before the transition? We can also ship over branches to NuRadioMC, but this is some work. Things would be easier if everything gets merged to the master before we make the transition.
Hi I will check soon. But there should be nothing major lying around.
NuRadioMC becomes cumbersome to debug/ test automatically if things change here and not there. There are many interdependencies. So from a NuRadioMC standpoint, it would be better to merge them, i.e. have a NuRadioReco folder in the NuRadioMC repository. Is there anyone, who has an objection to this? Maybe the folks who don't use the neutrino half of things, @fschlueter ?